Human Cytomegalovirus Genomes Survive Mitosis via the IE19Chromatin-Tethering Domain

Shelby M. Lyon,a,b Kristen D. Yetming,a,b Christina Paulus,c Michael Nevels,c Robert F. Kalejtaa,b

aInstitute for Molecular Virology, University of Wisconsin–Madison, Madison, Wisconsin, USAbMcArdle Laboratory for Cancer Research, University of Wisconsin–Madison, Madison, Wisconsin, USAcBiomedical Sciences Research Complex, University of St. Andrews, St. Andrews, United Kingdom

ABSTRACT The genomes of DNA tumor viruses regain nuclear localization after nu-clear envelope breakdown during mitosis through the action of a viral protein witha chromatin-tethering domain (CTD). Here, we report that the human cytomegalovi-rus (HCMV) genome is maintained during mitosis by the CTD of the viral IE19 pro-tein. Deletion of the IE19 CTD or disruption of the IE19 splice acceptor site reducedviral genome maintenance and progeny virion formation during infection of dividingfibroblasts, both of which were rescued by IE19 ectopic expression. The discovery ofa viral genome maintenance factor during productive infection provides new insightinto the mode of HCMV infection implicated in birth defects, organ transplant fail-ure, and cancer.

IMPORTANCE Human cytomegalovirus (HCMV) is the leading infectious cause ofbirth defects, represents a serious complication for immunocompromised HIV/AIDSand organ transplant patients, and contributes to both immunosenescence and car-diovascular diseases. HCMV is also implicated in cancers such as glioblastoma multi-forme (GBM) and infects ex vivo-cultured GBM tumor cells. In dividing tumor cells,the genomes of DNA tumor viruses regain nuclear localization after nuclear enve-lope breakdown during mitosis. This mitotic survival is mediated by a viral proteinwith a chromatin-tethering domain (CTD). Here, we report that the HCMV genome ismaintained in dividing fibroblasts by the CTD of the viral IE19 protein. The discoveryof a viral genome maintenance factor during productive infection could help explainviral genome dynamics within HCMV-positive tumors as well as during latency.

KEYWORDS cancer, chromatin, herpes, latency, mitosis, transcription

Viruses modify the host cell cycle to induce or synchronize cells in phases thatoptimally support infection (1). Human cytomegalovirus (HCMV) infection of pri-

mary human fibroblasts in the G0 or G1 phase drives them to the G1/S border butprevents them from progressing through the S phase (2–7). Under these conditions,productive infection initiates immediately and is highly efficient. The mechanisms ofthe rapid initiation of the productive cycle in G0/G1 cells and their synchronization atthe G1/S border are reasonably well understood and include the stimulation of viraltranscription by the viral pp71 protein, the stimulation of cell cycle progression by pp71and vCdk UL97, and the G1/S arrest instituted by the viral IE2 protein (4, 5, 8–11).

When HCMV infects cells already in the S phase, initiation of the productive cycle isdelayed until the infected cell doubles its DNA content, traverses the S and G2 phases,completes mitosis, and enters G0/G1 (12, 13). Productive replication in G2 is inefficientand in mitosis causes mitotic catastrophes resulting in abortive, nonproductive, cyto-cidal infections (14, 15). The mechanism of the delay in the productive replication cyclein S-phase cells involves the interaction of cyclin A2 with the viral pp150 protein tosuppress productive-phase viral transcription (16). As infected cells traverse mitosis,

Citation Lyon SM, Yetming KD, Paulus C,Nevels M, Kalejta RF. 2020. Humancytomegalovirus genomes survive mitosis viathe IE19 chromatin-tethering domain. mBio11:e02410-20. https://doi.org/10.1128/mBio.02410-20.

Editor Stacey Schultz-Cherry, St. JudeChildren's Research Hospital

Copyright © 2020 Lyon et al. This is an open-access article distributed under the terms ofthe Creative Commons Attribution 4.0International license.

Address correspondence to Robert F. Kalejta,rfkalejta@wisc.edu.

Received 24 August 2020Accepted 26 August 2020Published

RESEARCH ARTICLEMolecular Biology and Physiology

crossm

September/October 2020 Volume 11 Issue 5 e02410-20 ® mbio.asm.org 1

29 September 2020

on February 28, 2021 by guest

http://mbio.asm

.org/D

ownloaded from

cyclin A2 is naturally degraded, and immediate early (IE) transcription initiates in thesubsequent G0/G1 phase of the daughter cells, driving the delayed initiation of pro-ductive infection (15). Thus, HCMV devotes multiple mechanisms to synchronization ofinfected cells in favorable cell cycle phases while preventing cells that are activelyreplicating virus from entering mitosis, where mitotic catastrophes limit productivereplication (14, 15).

While the general mechanisms that HCMV uses to modulate the cell cycle have beensolved or are under study, the mechanisms that allow the virus infecting an S-phase cellto initiate productive replication after mitosis have largely been ignored. Duringmitosis, the nuclear envelope breaks down, and two independent, complete sets ofcellular chromosomes are pulled to opposite poles by kinetochore microtubules at-tached to chromosomal centromeres. New nuclear envelopes then reform around eachset of bundled chromosomes. Because viral genomes (including HCMV) lack a centro-mere, they cannot be reincorporated into daughter nuclei in this fashion. Early work,however, determined that essentially identical percentages of cells express IE proteinsat 24 h after HCMV infection regardless of whether the cells were infected in G0/G1 orS phase (12). This result strongly suggested that in S-phase-infected cells, the viralgenome is maintained in the nucleus through mitosis, yet no studies have investigatedmechanisms through which this process may occur during these delayed yet produc-tive infections.

For other viruses, mechanisms mediating the nuclear retention of their genomesduring mitosis are well established. Retroviruses integrate their genomes into hostchromosomes (17). The DNA tumor viruses human papillomavirus (HPV), Epstein-Barrvirus (EBV), and Kaposi’s sarcoma-associated herpesvirus (KSHV) maintain their extra-chromosomal genomes by using a virally encoded protein to tether them to cellularchromatin (18–20). The HPV E2, EBV EBNA1, and KSHV LANA proteins each containcarboxy-terminal domains that bind sequence specifically to their respective viralgenomes (21–25) and amino-terminal domains that bridge and tether the complex tocellular chromatin. The HPV E2 transactivation domain interacts with cellular chromatin-associated proteins (26–28), EBV EBNA1 domains A and B bind to AT-rich cellular DNAand cellular chromatin-associated proteins (29, 30), and the chromatin-binding domainof KSHV LANA binds histones H2A and H2B on cellular or viral genomes (31, 32).

The HCMV UL123 gene encodes a defined chromatin-tethering domain (CTD) at the3= end of its exon 4 (33, 34) that is dispensable for productive replication in asynchro-nous fibroblast cells (35, 36). In the context of the 72-kDa IE1 protein, the major proteinencoded by UL123, the CTD associates with the acidic patch of nucleosomes andmediates the colocalization of IE1 with condensed cellular chromosomes similar toKSHV LANA (37, 38). While the UL123 CTD has long been hypothesized as a viralgenome tether (33, 36, 37, 39, 40), the function of this domain has never been testedin mitotic cells.

Like the tethering proteins described above, IE1 dimerizes. The crystal structure ofthe core domain of rhesus CMV (RhCMV) IE1 (analogous to amino acids 27 to 379 of the491-amino-acid IE1 protein of HCMV) revealed an antiparallel dimer (41). Dimerizationof a truncated fragment of Myc epitope-tagged HCMV IE1 (amino acids 1 to 382) totruncated fragments of FLAG epitope-tagged HCMV IE1 (amino acids 1 to 382 or 1 to377) was observed by coimmunoprecipitation with Western blotting (41). In the crystalstructure, the dimerization interface encompassed the entire length of the core domainand was mainly hydrophilic. A subsequent report (42) used modeling based on theRhCMV IE1 structure to identify five residues within the core domain (K238, Q252, K300,L311, and R325) that provide the strongest effects on predicted binding affinity for theIE1 dimer. Based on alignment (43), the HCMV IE1 equivalent residues are K223, N237,N285, M296, and R310. How dimerization affects IE1 function, or whether otherUL123-encoded protein isoforms also dimerize, has not been explored.

EBV and KSHV CTDs support viral latency, and there is interest in the role of UL123exon 4 during HCMV latency (39). Unfortunately, no system is currently available toassay the maintenance of latent HCMV genomes through mitosis. Therefore, we tested

Lyon et al. ®

September/October 2020 Volume 11 Issue 5 e02410-20 mbio.asm.org 2

on February 28, 2021 by guest

http://mbio.asm

.org/D

ownloaded from

the requirement of the CTD for the only documented instance of the virus survivingmitosis: the infection of fibroblasts residing in the S phase of the cell cycle. We showthat the CTD, in the context of an alternatively spliced UL123 transcript encoding amonomeric protein isoform designated IE19, helps the viral genome to survive mitosisso that the viral genome can initiate the cascade of productive-phase gene expressionin daughter cells entering G1 phase after completing mitosis.

RESULTSS-phase-infected cells must pass through mitosis to support productive-phase

viral gene expression. We confirmed previous reports that G0/G1 fibroblasts infectedwith HCMV initiate IE gene expression quickly, while S-phase-infected fibroblastsinitiate IE gene expression only after a pronounced delay (12). Asynchronous, subcon-fluent fibroblasts were synchronized either in G0 by serum starvation or in S phase byaphidicolin, released from the block, and then infected (Fig. 1A). Released and mock-infected G0 cells entered mitosis after a substantial delay (�24 h), whereas infected cellsdid not (Fig. 1B), all as expected. Released S-phase cells entered mitosis much faster(�12 h), whether mock or HCMV infected (Fig. 1B), also as expected, indicating thevalidity of our synchronization and release protocols. HCMV-infected G0 cells were IE2positive as early as 4 h postinfection (hpi), the earliest time point examined, indicatingthat they initiated the productive replication cycle, and they achieved essentiallymaximum percentages of IE2-positive cells by 8 hpi (Fig. 1C). In contrast, HCMV-infectedS-phase cultures failed to show substantial percentages of IE2-positive cells as late as12 hpi, but by 24 hpi they had achieved a percentage of IE2-positive cells indistin-guishable from that of the G0 infected cultures (Fig. 1C). We conclude, like others beforeus, that IE gene expression is delayed in S-phase-infected cells until they traversemitosis and their daughter cells enter the subsequent G1 phase (12).

FIG 1 Human cytomegalovirus (HCMV) immediate early (IE) gene expression is silenced after S-phase infectionuntil cells pass through mitosis. (A) Experimental timeline for synchronization via aphidicolin treatment or serumstarvation, followed by infection and analysis. (B) Mitotic index was determined after media were supplementedwith Hoechst at 4 �g/ml and cells were imaged at the stated time points after release from synchronization. Thefraction of cells with condensed chromatin or split nuclei was calculated for at least 1,000 cells and is presentedas the mean � SD for each time point (n � 3). (C) Cells synchronized in G0 or S were infected at an MOI of 1 andwashed to release. Cells were fixed at the indicated time points for analysis by IF with antibodies for IE2 andstaining with Hoechst. Data are the means and SD for each time point (n � 3). (B and C) *, P � 0.05; **, P � 0.01;n.s., not significant (P � 0.1).

IE19 Helps the HCMV Genome through Mitosis ®

September/October 2020 Volume 11 Issue 5 e02410-20 mbio.asm.org 3

on February 28, 2021 by guest

http://mbio.asm

.org/D

ownloaded from

We next tested whether a lengthy delay is sufficient to make viral genomescompetent for transcription during an S-phase infection, or whether passage throughmitosis is required, by infecting S-phase cells synchronized with aphidicolin, releasingthem into either dimethyl sulfoxide (DMSO) (to permit cell cycle progression), aphidi-colin (to maintain S-phase arrest), or nocodazole (to synchronize cells in mitosis) for24 h, and monitoring IE2 expression (Fig. 2A). S-phase cells infected and released intoDMSO synthesized IE2 (Fig. 2B, lane 3), but those released into aphidicolin (Fig. 2B, lane5) or nocodazole (Fig. 2B, lane 6) did not. However, when infected S-phase cells werereleased into nocodazole for 24 h and then subsequently released from the nocodazoleblock for 6 h (to allow passage through mitosis), IE2 was expressed to a level indistin-guishable from that in infected cells directly released from aphidicolin into DMSO(Fig. 2B, compare lanes 4 and 3). Nocodazole itself had no effect on IE2 production inG0-infected cells that arrested at the G1/S border and never reached mitosis (Fig. 2B,compare lanes 1 and 2), indicating that the drug itself is not inhibitory to viral geneexpression unless it arrests cells in mitosis. From these experiments, we conclude thatS-phase-infected cells must pass through mitosis to gain competency for viral geneexpression in the subsequent G1 phase.

CTD-deficient HCMV displays reduced gene expression, genome levels, andproductive replication after S-phase infection compared to the wild type. S-phase-infected cells must pass through mitosis before initiating IE gene expression in thesubsequent G1 phase at levels equivalent to those in G0-infected cells. This suggeststhat viral genomes are present after mitosis at levels comparable to those prior tomitosis and therefore that viral genomes are retained in the nucleus during mitosis,perhaps by a dedicated mechanism. We tested if such a putative mechanism utilizedthe UL123 CTD. We created a CTD-deficient (ΔCTD) recombinant in the Towne strain ofHCMV that contains a stop codon at IE1 amino acid position 476, similar to onepreviously described (35). Our Towne ΔCTD recombinant produces a C-terminallytruncated IE1 protein smaller than full-length IE1 (Fig. 3A) that lacks the CTD and failsto associate with mitotic chromosomes (Fig. 3B) but grows with wild-type (WT) kineticsin G0 cells (Fig. 3C), all identical to the published virus (35).

We saw no differences in the fraction of IE2-positive cells between WT and the ΔCTDvirus during G0 infection (Fig. 4A to D). However, the ΔCTD virus displayed a signifi-cantly reduced fraction of IE2-positive cells compared to WT virus after S-phaseinfection (Fig. 4A to D). The magnitude of the defect was exacerbated at a 3.3-fold-

FIG 2 S-phase cells infected with HCMV must pass through mitosis to gain competency for IE gene expression. (A)Experimental timeline for the six infection conditions in panel B. Cells were synchronized with either aphidicolin treatment orserum starvation, followed by infection and supplementation with media containing DMSO, nocodazole, or aphidicolin for 24h. Cells were fixed at 24 hpi for IE2 and Hoechst analysis by IF. In condition 4, cells were washed at 24 hpi and released for6 h (30 hpi) before fixing. (B) Fraction of IE2-positive cells was calculated for each condition and is presented as the mean andSD (n � 3). ***, P � 0.001; n.s., not significant (P � 0.1).

Lyon et al. ®

September/October 2020 Volume 11 Issue 5 e02410-20 mbio.asm.org 4

on February 28, 2021 by guest

http://mbio.asm

.org/D

ownloaded from

lower multiplicity of infection (Fig. 4C), and the defect was observed using twoindependent methods to synchronize cells in the S phase: aphidicolin (Fig. 4A to C) orcontact inhibition and release (Fig. 4D). Identical results were obtained when produc-tion of green fluorescent protein (GFP), encoded on a transgene, by each virus wasused to score viral gene expression (Fig. 4A, C, and D). Transcription of the GFPtransgene in HCMV is dependent on IE protein synthesis (44), and therefore, GFPprotein accumulation was assayed 24 h after IE2 to ensure that all cells had theopportunity to express GFP. We conclude that in S-phase-infected cells, the UL123 CTDenhances viral gene expression in the subsequent G1 phase after completion of mitosis.

The CTD plays no apparent role in viral gene expression in G0 but significantlyimpacts viral gene expression in cells that pass through mitosis (Fig. 4A), whereextrachromosomal DNAs can be lost from the nucleus because of nuclear envelopebreakdown (45–47). Therefore, we asked if the CTD helps to maintain viral genomelevels in S-phase-infected cells that pass through mitosis. WT virus showed no differ-ences in genome levels between G0- and S-phase cells at 6 hpi (Fig. 5A), indicating thatthe virus entered G0- and S-phase cells equivalently, or at 24 hpi (Fig. 5B), indicatingthat viral genomes that passed through mitosis (S-phase infection) and those that didnot (G0-phase infection) were maintained equivalently. The ΔCTD virus showed nodifferences in genome levels at 6 hpi between G0- and S-phase cells (Fig. 5A), indicatingthat this mutant virus also entered G0- and S-phase cells equivalently. However, theΔCTD virus showed lower levels of viral genomes at 24 hpi after S-phase infection thanG0-phase infection (Fig. 5B). The ΔCTD virus delivered more genomes to cells than didthe WT virus (Fig. 5A). There is precedent for a viral mutant to display no growth defectbut a higher genome/PFU ratio than WT virus (48). Despite the delivery of �3.5-fold

FIG 3 Characterization of CTD-deficient IE1 mutant Towne virus. MRC-5 cells were infected withwild-type (WT) or CTD-deficient (ΔCTD) virus. (A) At the indicated times postinfection, whole-cell extractswere prepared and subjected to Western blot analysis with antibodies against HCMV IE1 and IE2(MAB810R) and cellular GAPDH. (B) At 48 hpi, cells were fixed and stained with anti-cytomegalovirus IE1antibody (ab30924) and DAPI (4=,6-diamidino-2-phenylindole). Representative images are shown ingrayscale with a color merge. (C) Every 24 h, viral replication was assessed by qPCR-based relativequantification of HCMV DNA from culture supernatants with primers specific for the viral UL54 promotersequence. Data are means � SD (n � 3).

IE19 Helps the HCMV Genome through Mitosis ®

September/October 2020 Volume 11 Issue 5 e02410-20 mbio.asm.org 5

on February 28, 2021 by guest

http://mbio.asm

.org/D

ownloaded from

more genomes than WT virus to S-phase cells, the ΔCTD virus showed �2.5-fold fewergenomes than WT virus after the cells passed through mitosis. Our data indicate thatΔCTD genomes that passed through mitosis were lost more frequently than WTgenomes.

FIG 4 HCMV gene expression is reduced after S-phase infection with a CTD-deficient virus. (A) Thefraction of IE2-positive cells was calculated at 24 hpi and the fraction of GFP-positive cells was calculatedat 48 hpi following G0- or S-phase infection at an MOI of 1. Data are mean IE- and GFP-positive ΔCTDfractions relative to WT fractions � SD (n � 3). (B) Cells synchronized in G0 or S were infected at an MOIof 1 and washed to release. Cells were fixed at 24 hpi for analysis by IF with antibodies for IE2 andstaining with Hoechst. Representative images are provided in grayscale alongside Hoechst and a colormerge. n � 3. (C and D) IE2- and GFP-positive fractions were calculated at 24 h and 48 h, respectively,after G0- or S-phase infection at an MOI of 0.3 (C) and after S-phase synchronization by contact inhibitionand release instead of aphidicolin treatment (D). Data are mean IE- and GFP-positive ΔCTD fractionsrelative to WT fractions, with SD (n � 3). (A, C, and D) *, P � 0.05; **, P � 0.01; n.s., not significant (P � 0.1).

Lyon et al. ®

September/October 2020 Volume 11 Issue 5 e02410-20 mbio.asm.org 6

on February 28, 2021 by guest

http://mbio.asm

.org/D

ownloaded from

The defects in productive-phase transcription and genome level maintenance ob-served when ΔCTD virus infects S-phase cells extended to defects in progeny virionformation. In G0 cells, WT and ΔCTD virus produced infectious progeny in synchronyand to identical levels (Fig. 5C). In S-phase cells, WT virus showed a delay in progenyvirion formation (Fig. 5C) similar to the delay observed in viral gene expression (Fig. 1C),but the overall output of infectious progeny eventually matched the level achieved inG0 cells. The ΔCTD virus, however, displayed a more pronounced delay in progeny virusformation in S-phase-infected cells, and the overall output of infectious progeny wasreduced �100-fold from the level achieved in G0 cells (Fig. 5C). We conclude that theΔCTD virus shows pronounced defects in viral gene expression, genome levels, andproductive replication after infection of S-phase cells because its genome is lost as theinfected cells pass through mitosis.

An alternatively spliced UL123 transcript encoding the CTD-containing IE19isoform is upregulated after S-phase infection. UL123 must be transcribed duringS-phase infections, because the UL123 CTD is required for genome survival duringmitosis. The UL123 gene produces transcripts encoding IE1, IE1x4, and IE19 (39, 49–51).We utilized RNA ligase-mediated rapid amplification of cDNA ends (RLM-RACE) withprimers anchored in the CTD to identify UL123 transcripts produced following S-phaseinfection (Fig. 6A). CTD-encoding 1.4-kb and 0.4-kb transcripts consistent with thoseencoding IE1 and IE19, respectively, were detected (Fig. 6B). We sequenced 10 inde-pendent clones derived from the RLM-RACE reaction of S-phase-infected cells andfound that two encoded the canonical IE1 transcript while 8 encoded the IE19 tran-script (Fig. 6C). No other UL123 transcripts were identified. For comparison, we se-quenced 10 independent clones derived from RLM-RACE reactions from G0-infectedcells, and all 10 represented the canonical IE1 transcript (Fig. 6C).

To better quantify expression of individual transcript species during infection, weemployed quantitative reverse transcription-PCR (qRT-PCR) primer sets specific for IE1or IE19 that spanned their unique exon 3– exon 4 splice junctions. We also used aprimer set in the 5= end of exon 4 that would detect (but not differentiate between)transcripts encoding IE1 and IE1x4 but would not detect those encoding IE19. Consis-tent with S-phase restriction of the full cascade of productive-phase gene expression(12), IE1 transcript levels were reduced in S-phase cells compared to G0 cells (Fig. 6D).Equivalent results with the IE1 and IE1/IE1x4 primers suggested that a transcriptcarrying exclusively UL123 exon 4 was not expressed. IE19 transcripts showed theopposite trend, being higher in S-phase cells than G0 cells (Fig. 6D). Importantly, theratio of IE19 to IE1 in S-phase cells was 10-fold higher than in G0 cells (Fig. 6E). Ourresults indicate that transcripts encoding IE19 are significantly more abundant in

FIG 5 HCMV genome levels and productive infection are reduced after S-phase infection with a CTD-deficient virus. (A)Cells synchronized in G0 or S phase were infected for 6 h and harvested for analysis of viral DNA (UL123 exon 3) by qPCR.Data were normalized to cellular DNA (GAPDH) and WT G0 and are presented as means and SD (n � 3). (B) Cellssynchronized in G0 or S phase were harvested at 24 hpi for analysis of viral DNA (UL123 exon 3) by qPCR. Data werenormalized to GAPDH and input genomes (6 hpi) and are presented as means and SD (n � 3). (C) Growth curves wereconstructed following G0- or S-phase infection with the indicated viruses at an MOI of 0.1. Cell-free virus was collected atthe stated days postinfection for titration by standard plaque assay. Data are means � SD (n � 3). (A and B) n.s., notsignificant (P � 0.1).

IE19 Helps the HCMV Genome through Mitosis ®

September/October 2020 Volume 11 Issue 5 e02410-20 mbio.asm.org 7

on February 28, 2021 by guest

http://mbio.asm

.org/D

ownloaded from

S-phase cells than G0 cells and implicate IE19 as the CTD-containing protein with thepotential to mediate nuclear genome maintenance during S-phase infections.

IE19-deficient HCMV displays reduced gene expression, genome levels, andproductive replication after S-phase infection compared to the wild type. Wetested if IE19 contributed to the nuclear retention of viral genomes during mitosis usingan IE19-deficient recombinant in the Towne strain of HCMV (WTSS) containing a silentmutation in the 3= splice site of IE19 (50). In G0- or S-phase cells, the WTSS virus failedto generate IE19-encoding mRNAs (Fig. 7A). Both the WTSS and the ΔCTD virusesdisplayed significantly reduced fractions of IE2-positive cells compared to WT virus afterS-phase infection (Fig. 7B). Identical results were obtained when production of GFPencoded on a transgene by each virus was used to score viral gene expression (Fig. 7C).WT virus showed no differences in genome levels at 6 or 24 hpi between G0- andS-phase cells, and the ΔCTD and WTSS viruses showed no differences in genome levelsat 6 hpi between G0- and S-phase cells. However, both mutant viruses showed lowerlevels of viral genomes at 24 hpi after S-phase infection than G0-phase infection(Fig. 7D), indicating that ΔCTD and WTSS genomes that passed through mitosis werelost more frequently than WT genomes.

The defects in productive-phase transcription and genome level maintenance ob-served when ΔCTD or WTSS virus infected S-phase cells extended to defects in progenyvirion production. In G0 cells, WT, ΔCTD, and WTSS viruses produced infectious progenyto identical levels (Fig. 7E). In S-phase cells, the overall output of infectious progeny forWT virus matched the level achieved in G0 cells but was reduced 10-fold for the ΔCTDand WTSS viruses (Fig. 7E). We conclude that the WTSS virus, similarly to the ΔCTD virus,

FIG 6 An alternatively spliced UL123 transcript encoding the CTD-containing IE19 protein is upregulated after S-phaseinfection. (A) Schematic of the MIE locus showing the splicing of the CTD-containing transcripts IE1, IE1x4, and IE19. Primersused for RLM-RACE are indicated by arrows, and the predicted product size for each transcript is presented on the right.(B) RLM-RACE was conducted on RNA isolated from G0- or S-phase cells infected at an MOI of 3 at 6 hpi. The image showsnested PCR products run on a 1% agarose gel and stained with ethidium bromide. Lane 1 is the CTD products from theG0-phase sample, and lane 2 is the CTD products from the S-phase sample. (C) Original PCRs for G0- and S-phase RLM-RACEsamples were cloned. Ten colonies were picked from each and sequenced from the forward and reverse directions tocapture the full-length read. All sequences were identical to either IE1 or IE19, and the number of each is presented. (D)Cells synchronized in either G0 or S phase and infected at an MOI of 1 were harvested at 6 hpi. RNA was analyzed byqRT-PCR for the indicated transcripts. Data are mean RNA levels relative to cellular GAPDH and SD (n � 3). *, P � 0.05. (E)The ratio of IE19 transcripts to IE1 transcripts was calculated from the qRT-PCR analysis of RNA from G0- or S-phaseinfections at 6 hpi.

Lyon et al. ®

September/October 2020 Volume 11 Issue 5 e02410-20 mbio.asm.org 8

on February 28, 2021 by guest

http://mbio.asm

.org/D

ownloaded from

shows pronounced defects in viral gene expression, genome levels, and productivereplication after infection of S-phase cells because its genome is lost as the infectedcells pass through mitosis in the absence of the CTD of the IE19 protein. Therefore, IE19,with its CTD, facilitates the retention of viral genomes in dividing cells.

IE19, but not IE1 or IE1x4, complements the deficiencies in gene expression,genome levels, and productive replication displayed by CTD- or IE19-deficientviruses after S-phase infection. Finally, we used a complementing cell line approachto determine if IE1, IE1x4, or IE19 expressed in trans could complement the S-phaseinfection defects observed with the ΔCTD and WTSS viruses. Primary normal humandermal fibroblasts (NHDFs) were transduced with lentiviruses encoding N-terminallytagged cDNAs for either IE1, IE1x4, IE19, or IE19ΔCTD (IE19 missing the CTD) and apuromycin resistance gene. Puromycin-resistant populations were selected that ex-pressed the appropriately sized protein encoded by the transduced UL123 cDNA(Fig. 8A). IE1 migrated as the expected 72-kDa species. IE1x4 migrated at �55 kDa,which is comparable to the �60-kDa species previously reported (39). IE19 migrated asthe expected �38-kDa species even though its predicted molecular weight is 19 kDa(51), and IE19ΔCTD migrated at a molecular weight consistent with the artificialtruncation. The 38-kDa apparent molecular weight of IE19 is exactly twice its predictedmolecular weight, and the DNA tumor virus CTD-containing mitotic maintenanceproteins HPV E2, EBV EBNA1, and KSHV LANA all function as dimers (52–56). However,in experiments where we detected dimerization of full-length IE1, we were unable to

FIG 7 IE19-deficient HCMV displays reduced gene expression, genome levels, and productive replication after S-phase infection. (A) RNAharvested from either G0- or S-phase infections with WT or WTSS Towne at 24 hpi was analyzed by qRT-PCR for IE19. Data are mean RNA levelsrelative to cellular GAPDH and SD (n � 3). (B) The fraction of HCMV-positive cells after infection in G0 or S phase with the indicated viruses at anMOI of 1 was measured by IF for IE2 and Hoechst at 24 hpi. Data are means and SD (n � 3). (C) The fraction of HCMV-positive cells after infectionin G0 or S phase with the indicated viruses at an MOI of 1 was measured by live-cell imaging for GFP and Hoechst at 48 hpi. Data are means andSD (n � 3). (D) DNA harvested at 24 hpi from G0- or S-phase infections with the indicated viruses at an MOI of 1 was analyzed by qPCR for viralDNA (UL123 exon 3). Data are presented as mean DNA levels relative to cellular GAPDH, with SD (n � 3). (E) Cell-free virus was collected at 8 dpi(days postinfection) after G0- or S-phase infection with the indicated viruses at an MOI of 0.1. Endpoint titers were determined by standard plaqueassay. Data are means and SD (n � 3). (B to D) *, P � 0.05; **, P � 0.01; n.s., not significant (P � 0.1).

IE19 Helps the HCMV Genome through Mitosis ®

September/October 2020 Volume 11 Issue 5 e02410-20 mbio.asm.org 9

on February 28, 2021 by guest

http://mbio.asm

.org/D

ownloaded from

detect IE19 dimerization (Fig. 8B and C). Differentially epitope (FLAG or HA)-tagged IE1coimmunoprecipitated the oppositely tagged protein (HA or FLAG, respectively) intransfection assays (Fig. 8B) in an efficient manner (Fig. 8C), but no association wasdetected between differentially tagged forms of IE19 (Fig. 8B and C). Thus, it is likelythat the acidic nature of the amino acid sequence of IE19 causes aberrant migration onSDS-PAGE, as has been detected for other proteins (57–59).

The IE1, IE19, and IE19ΔCTD proteins displayed a diffuse nuclear localization ininterphase cells (Fig. 8D), consistent with the presence of the known nuclear localiza-tion signal (NLS) encoded within the exon 2 sequences of each cDNA (34). The IE1x4protein, whose subcellular localization was not analyzed when its existence wasreported (39), showed diffuse cytoplasmic staining in interphase cells (Fig. 8D), consis-tent with the lack of the NLS sequences encoded in exon 2. In mitotic cells, IE1 and IE19colocalized with cellular chromatin (Fig. 8E), consistent with their identical encodedCTDs. Neither IE19ΔCTD nor IE1x4 colocalized with cellular chromatin in mitotic cells(Fig. 8E), consistent with IE19ΔCTD lacking the CTD and IEx4 localizing to the cytoplasm.

Both the ΔCTD and WTSS viruses displayed a significantly reduced fraction ofIE2-positive cells compared to WT virus after S-phase infection of the parental NHDF

FIG 8 Characterization of fibroblasts expressing IE1, IE1x4, IE19, or IE19ΔCTD. (A) Lysates from equal numbers of cells expressing the indicated protein wereharvested and analyzed with the indicated antibodies by Western blotting. Approximate sizes are shown. n � 3. (B) Transfected 293T cells expressing 3�FLAG-or HA-tagged IE19 and IE1 proteins were cross-linked with 1% formaldehyde, and lysates were subjected to immunoprecipitation with the indicated antibody,separation by SDS-PAGE, and analysis by Western blotting with antibodies against the FLAG and HA epitopes. n � 3. (C) The ratios of signal intensities of HAIP/FLAG IP for the FLAG blots and FLAG IP/HA IP for the HA blots were calculated for IE1 and IE19. Error bars indicate standard deviations. n � 3. ***, P � 0.001.(D) Cells expressing the indicated proteins were stained for IE1, FLAG, and Hoechst by IF. Representative images are presented in grayscale with a color merge.n � 3. (E) Cells expressing the indicated proteins were synchronized with aphidicolin for 24 h and washed to release. At 12 h postrelease, cells were fixed andstained for IE1, FLAG, and Hoechst by IF. Representative images are presented in grayscale with a color merge. n � 3.

Lyon et al. ®

September/October 2020 Volume 11 Issue 5 e02410-20 mbio.asm.org 10

on February 28, 2021 by guest

http://mbio.asm

.org/D

ownloaded from

cells, as well as after S-phase infection of NHDFs expressing IE1, IE1x4, or IE19ΔCTD(Fig. 9A), indicating that none of these proteins complemented the defects inherent inthe absence of the CTD (ΔCTD virus) or the absence of the IE19 protein (WTSS virus).Similarly, neither IE1, IE1x4, nor IE19ΔCTD was able to complement S-phase-infectiondefects in GFP expression (Fig. 9B), viral DNA maintenance (Fig. 9C), or infectious-virionproduction (Fig. 9D) for the ΔCTD or WTSS viruses. However, all known S-phase-infection defects for ΔCTD and WTSS viruses were complemented in IE19-expressingcells. Both the ΔCTD and WTSS viruses displayed IE2 expression (Fig. 9A), GFP expres-sion (Fig. 9B), viral genome maintenance (Fig. 9C), and infectious-progeny production(Fig. 9D) indistinguishable from those of WT virus after infection of IE19-expressingfibroblasts, indicating that IE19 expressed in trans complements the defects inherent inthe absence of the CTD (ΔCTD virus) or the absence of the IE19 protein (WTSS virus).In total, we conclude that HCMV genomes are maintained through mitosis in produc-tively infected fibroblasts by the CTD of IE19.

DISCUSSION

Even though they contain identical CTDs, IE1 does not mediate viral genome mitoticmaintenance during productive infection of S-phase cells, while IE19 does (Fig. 9). Asidefrom the missing amino acids, another difference between the two proteins is that IE1is detected as a dimer and IE19 is not (Fig. 8B and C). IE1 dimerizes through its coredomain (41, 42). While the beginning of the core domain (amino acids 25 to 85) isencoded by exon 2 and thus is present in IE19, the majority of the core domain (amino

FIG 9 IE19, but not IE1 or IE1x4, complements the deficiencies in gene expression, genome levels, andproductive replication displayed by CTD- or IE19-deficient viruses after S-phase infection. (A) Cellsexpressing the indicated proteins were synchronized in G0 or S phase and infected at an MOI of 1 withthe indicated viruses. The fraction of HCMV-positive cells was measured by IF for IE2 and Hoechst at 24hpi. Data are mean fractions of positive cells from S-phase infection relative to G0 infection, with SD(n � 3). (B) Cells expressing the indicated proteins were synchronized in G0 or S phase and infected atan MOI of 1 with the indicated viruses. The fraction of HCMV-positive cells was measured by live-cellimaging for GFP and Hoechst at 48 hpi. (C) DNA harvested at 24 hpi from G0- or S-phase infection at anMOI of 1 was analyzed by qPCR for viral DNA (UL123 exon 3) and cellular DNA (GAPDH). Data are meanDNA levels normalized to GAPDH from S-phase infection relative to G0 infection, with SD (n � 3). (D)Cell-free virus was collected at 8 dpi after G0- or S-phase infection with the indicated viruses at an MOIof 0.1 Endpoint titers were determined by standard plaque assay. Data are mean S phase infection titersrelative to G0 infection, with SD (n � 3). (B to D) *, P � 0.05; **, P � 0.01; ***, P � 0.001; n.s., not significant(P � 0.1).

IE19 Helps the HCMV Genome through Mitosis ®

September/October 2020 Volume 11 Issue 5 e02410-20 mbio.asm.org 11

on February 28, 2021 by guest

http://mbio.asm

.org/D

ownloaded from

acids 86 to 379) is encoded in the region of exon 4 that is spliced out to make theIE19-encoding mRNA and therefore is not found in the IE19 protein. Of the five residuespredicted to drive IE1 dimerization, none is present in IE19. Thus, IE19 would not bepredicted to, nor did we observe it to, dimerize (Fig. 8B).

Tethering of viral genomes or chromatin to cellular chromatin requires the tetheringprotein to simultaneously bind both nucleic acid complexes either directly or indirectly.IE19 lacking the CTD does not associate with cellular chromatin (Fig. 8D), implicatingthe CTD as the functional domain that might bind to cellular chromatin during themitotic maintenance of viral genomes. The fact that IE19 is not a dimer and thereforefunctions with only a single CTD implies that an additional region of the protein otherthan the CTD is responsible for binding viral DNA directly or indirectly. A postulatedviral DNA-associating region could be the acidic domain, as has been speculated forIE1x4 (39), or it could be a different region of the IE19 protein. Perhaps IE1 dimerization,or the folded nature of the core domain, sterically blocks the postulated but as-yet-unidentified functional surface required for viral DNA association. Thus, while IE1retains the CTD and binds chromatin, perhaps it does not bind viral DNA and as suchcannot mediate viral genome maintenance through mitosis.

In addition to its inability to shepherd viral genomes through mitosis, IE1 is a highlyimmunogenic protein. During the extended nonproductive phase of an S-phase infec-tion where viral immune evasion genes do not appear to be expressed, avoidingimmune detection may be paramount to viral survival in vivo. Interestingly, eight of the10 most prominently recognized CTL epitopes in IE1 (60) are absent in IE19. Along withalterations in the presence or sequence of CTL epitopes, maintaining low viral proteinlevels is an effective immune evasion strategy common to many viruses. Thus, andperhaps not surprisingly, like others before us (49, 50), we have been unable to detectthe IE19 protein in HCMV-infected cells, despite numerous attempts with multipledifferent protein enrichment or stabilization protocols. In the absence of a detectableprotein, it remains formally possible that the IE19-coding mRNA functions in a non-coding capacity. However, because all other known examples of viral genome main-tenance through mitosis are mediated by viral proteins, we hypothesize that the IE19protein mediates HCMV genome maintenance through mitosis, despite our inability todetect it with the currently available reagents. We suspect that antibodies with higheraffinity or greater specificity would permit detection of the IE19 protein during HCMVinfection and that IE19 levels are simply kept as low as possible to avoid immunedetection and clearance of the infected cell. Thus, the use of the IE19 protein for viralgenome survival through mitosis during S-phase infection, as opposed to largerCTD-containing proteins encoded by UL123 like IE1 or IEx4, may have emerged not onlyfor functional competency but also to escape immune surveillance.

Viruses other than HCMV for which CTD-mediated extrachromosomal viral genomemaintenance during mitosis has been demonstrated are oncogenic. HCMV has beenimplicated as a cofactor for glioblastoma multiforme (GBM) tumors (61–63). During exvivo infections of primary GBM samples, HCMV productively replicates in some cells, butin others it increases their capacity for self-renewal (64), indicating perhaps that in thesedividing cells the virus is being maintained in the absence of progeny virion formation.A mechanism for genome maintenance during mitosis would appear to be required tomaintain this type of nonproductive infection, similar to how HPV, EBV, and KSHVgenomes are maintained in tumor cells by chromatin tethering in the absence ofprogeny virion formation. In addition to cancers, HCMV is also implicated in prolifera-tive atherosclerotic diseases such as transplant-associated vasculopathy (65–67). Ourdata indicate that IE19 activity facilitates the infection of dividing cells by providing agenome retention function during mitosis. Should an inhibitor of IE19 function emerge,it could impede HCMV infection in dividing cells and as such might be useful incombination therapies for the treatment of diseases such as cancers and cardiovasculardiseases that may be potentiated by HCMV infection of proliferating cells.

The mitotic maintenance of DNA tumor virus genomes occurs in cells that undergomultiple mitoses, and therefore, the extrachromosomal viral genomes replicate to

Lyon et al. ®

September/October 2020 Volume 11 Issue 5 e02410-20 mbio.asm.org 12

on February 28, 2021 by guest

http://mbio.asm

.org/D

ownloaded from

prevent the dilution of the viral genome that would eventually limit the number of viralgenome-positive cells (18, 19). The mitotic maintenance of HCMV genomes occurs inS-phase cells that undergo only one mitosis, after which productive replication initiates,arrests the cell cycle, and eventually kills the cell. Thus, in the absence of a need forinfected-cell expansion, we have yet to examine whether the HCMV genome replicatesafter S-phase infection of fibroblasts.

The mitotic maintenance of DNA tumor virus genomes occurs during nonproductive(persistent or latent) infections. Our demonstration here that HCMV genomes inproductively infected cells survive mitosis through a mechanism similar to thoseutilized by the DNA tumor viruses leads one to consider if such an event also transpiresduring latency. Currently, it is unclear whether cells latently infected with HCMV dividein vivo or in vitro, and there are no assays that quantitate latent genome replication orsurvival through mitosis. However, our demonstration here of a dedicated, IE19-dependent mechanism for viral genome survival through mitosis in differentiated cellsmakes it possible that an identical or similar mechanism maintains latent genomes ina dividing population of the incompletely differentiated myeloid cells where HCMVmaintains latency.

CD34� hematopoietic progenitor cells infected with an HCMV recombinant lackingexon 4 (and therefore lacking IE19 and the CTD) were found to harbor less viral DNA35 days after infection than WT virus-infected cells (39). However, that study did notquantitate the viral DNA, did not test for actual viral genome replication or mainte-nance in infected cells that were dividing, and did not control for the potentialproductive amplification of the WT virus used (the exon 4 mutant is unable to replicatein noncomplementing cell lines and thus could not productively amplify in theseexperiments).

The major immediate early promoter (MIEP) that directs UL123 transcription isrepressed during HCMV latency (68). Occasionally, transcripts annotated to UL123 aredetected during latency that may be explained by nascent reactivation events, tran-scriptional noise, or overly high-MOI in vitro infections. However, they may alsorepresent CTD-containing transcripts (possibly IE19) making proteins that mediate viralgenome replication and/or mitotic maintenance during latency. More work is neededto determine whether IE19 or some other CTD-containing protein mediates viralgenome maintenance during HCMV latency.

Finally, the repression of IE1 expression after S-phase infection and its activation inthe subsequent G1 phase mimic, to some degree, the silencing of IE1 during theestablishment and maintenance of latency and the animation of IE1 expression duringreactivation to productive replication (69). It remains to be determined whether similaror different mechanisms control UL123 transcriptional suppression, activation, andsplicing during S phase and latent infections.

MATERIALS AND METHODSCells and infections. Normal human dermal fibroblasts (NHDFs; Clonetics), human embryonic lung

fibroblasts (MRC5; ATCC), and 293T cells were maintained in Dulbecco’s modified Eagle medium (DMEM;Sigma) supplemented with 10% (vol/vol) fetal bovine serum (FBS; Sigma) and 1� penicillin-streptomycinwith L-glutamine (PSG) (G1146; Sigma). Synchronization by serum starvation was conducted by platingcells at 1 � 104 cells/cm2 for 16 h. Cells were washed three times with Dulbecco’s phosphate-bufferedsaline (DPBS; Invitrogen) and incubated for 24 h in low-serum medium (0.1% FBS–DMEM�PSG) beforeinfection. Synchronization by contact inhibition was conducted by feeding confluent cells completemedium daily for 120 h, followed by replating at 2 � 104 cells/cm2 for 16 h before infection. Synchro-nization by aphidicolin treatment was conducted by plating cells at 1 � 104 cells/cm2 for 16 h andchanging the medium to complete DMEM containing 2 �g/ml aphidicolin for 24 h before infection.Infections with HCMV were conducted in minimal volume for 60 min at 37°C with rocking every 10 min,followed by the addition of fresh medium to normal volume conditions. A multiplicity of infection (MOI)of 1 was used in all experiments unless otherwise specified to avoid the spurious retention of extra-chromosomal DNAs (unaided by a viral retention function) that occurs in direct relation to their numberswithin individual nuclei. Furthermore, an MOI of 1 generates �50% IE-positive cells during G0 infections,easily allowing either increases or decreases of IE-positive cells to be quantitated. WT, ΔCTD, and WTSSviruses are in the Towne backbone and express GFP from a simian virus 40 (SV40) promoter. In the ΔCTDvirus, a single engineered point mutation converts codon 476 from a glycine codon (GGA) to a stopcodon (TGA). The identity of the viruses was confirmed by sequencing of the UL123 region. The WTSS

IE19 Helps the HCMV Genome through Mitosis ®

September/October 2020 Volume 11 Issue 5 e02410-20 mbio.asm.org 13

on February 28, 2021 by guest

http://mbio.asm

.org/D

ownloaded from

virus was provided by Jeff Meier and Mark Stinski (University of Iowa) and was described previously (50).Two engineered point mutations in the WTSS virus convert the splice acceptor site within exon 4 thatis required to generate IE19 from 5=-CAG AGT-3= to 5=-CgG AaT-3=. These mutations inhibit splicing butdo not affect the amino acid sequence of IE1. The identity of the viruses was confirmed by demonstratingthat the WT produces the spliced mRNA encoding IE19 and that WTSS does not (Fig. 7A).

Inhibitors and antibodies. Where indicated in the figure legends and on the figures, aphidicolin(2 �g/ml; Sigma) or nocodazole (50 ng/ml; Sigma) was diluted in fresh medium and added to cells. Thefollowing antibodies were from commercial sources: anti-glyceraldehyde-3-phosphate dehydrogenase(GAPDH) (ab9485; Abcam), anti-tubulin (DM 1A; Sigma), anti-FLAG (PA1-984B, Thermo), anti-FLAG (M2;Sigma), anti-HA (HA.11, BioLegend), anti-cytomegalovirus for IE1 (ab30924; Abcam), and anti-cytomegalovirus for IE1 and IE2 (MAB810R; Millipore). Monoclonal antibodies against IE1 (1B12) and IE2(3H9) were described previously (70). Infrared (IR) dye 680- and 800-conjugated secondary antibodies(Li-Cor) were used for Western blotting. Alexa Fluor 488-conjugated secondary antibody (catalog no.A-11017; Invitrogen) and Alexa Fluor 594-conjugated secondary antibody (catalog no. A-11020; Invitro-gen) were used for immunofluorescence.

Transductions. Lentiviral transduction of fibroblasts was performed as previously described (71).Fibroblasts were transduced with pSin lentiviral vectors expressing IE1, IE19, IE1x4, or IE19ΔCTD, all with3�FLAG tags on their N termini. IE19ΔCTD was created by truncating IE19 at amino acid 156 andinserting a stop codon. Transduced cells were maintained in complete medium containing 1 �g/mlpuromycin (Sigma).

Coimmunoprecipitations. 293T cells were cotransfected with pSG5 expression vectors encoding3�FLAG- or HA-tagged IE19 and IE1 alleles using Lipofectamine 2000 (Life Technologies), according tothe manufacturer’s instructions. At 2 days posttransfection, the cells were harvested and cross-linkedwith 1% formaldehyde for 10 min at room temperature. To quench, glycine was added to a finalconcentration of 125 mM and incubated for 5 min at room temperature. The cross-linked cell lysates werepelleted, washed with cold PBS, and resuspended in 1� cell lysis buffer (Cell Signaling Technologies)supplemented with 1 mM phenylmethylsulfonyl fluoride (PMSF; Cell Signaling Technologies). Sampleswere sonicated three times on ice for 5 s each and centrifuged at 14,000 � g for 10 min at 4°C, and thesupernatant (cell lysate) was transferred to a new tube. Cell lysates were incubated with rotationovernight with 2 �g of anti-FLAG (M2), anti-HA.11, or mouse IgG isotype control antibody (I5381; Sigma).Protein A/G magnetic beads (Thermo Scientific) were prewashed with 1� cell lysis buffer plus 1 mMPMSF and then incubated with the immunocomplex for 40 min at room temperature with rotation. Thebeads were pelleted using a magnetic rack and washed five times with 1� cell lysis buffer plus 1 mMPMSF. For SDS-PAGE, 2� SDS protein sample buffer (125 mM Tris-HCl [pH 6.8], 4% SDS, 20% glycerol,0.004% bromophenol blue) supplemented with 5% �-mercaptoethanol was added, and the sampleswere boiled for 10 min prior to separation by SDS-PAGE and analysis by Western blotting. Signalintensities were acquired using Image Studio version 5.2 software (Li-Cor). The ratios of signal intensitiesof HA IP/FLAG IP for the FLAG blots and FLAG IP/HA IP for the HA blots are presented as means withstandard deviations of three biological replicates. An unpaired, two-tailed Student’s t test was used todetermine the statistical difference between IE1 and IE19 (P � 0.001).

Indirect immunofluorescence assays. Cells were cultured, synchronized, and/or infected for indi-rect immunofluorescence on a �-Slide VI 0.4 (80606; Ibidi). At various time points, cells were washedtwice with cold PBS and fixed with 1% paraformaldehyde (PFA) either for 30 min at room temperatureor overnight at 4°C. Immunofluorescence (IF) was assessed as previously described (10) except to followmanufacturer’s recommendations (AN03 and MV18; Ibidi) for volumes and handling. Images werecollected with a Nikon Ti-Eclipse inverted wide-field microscope, taken with a CoolSnap HQ camera, andrecorded with Nikon NIS Elements software (v 4.00.03). At least 500 nuclei were counted per condition.Images were processed and background subtracted using NIH FIJI/ImageJ software.

DNA and mRNA analysis. For Fig. 3C, viral DNA was isolated from culture supernatants with theDNeasy blood and tissue kit (69506; Qiagen) and quantitated as previously described (72). For all otherexperiments, total DNA was isolated using a genomic DNA minikit (IB47202; IBI). Total DNA (200 ng) wasanalyzed by quantitative PCR using iTaq Universal SYBR green Supermix (172-5124; Bio-Rad) on anABI7900HT real-time PCR system (Applied Biosystems) instrument. Viral genomes were amplified withprimers specific to exon 3 of IE1/IE2 (73) or the UL54 promoter region (74), and cellular DNA wasamplified with glyceraldehyde-3-phosphate dehydrogenase (GAPDH) primers (75). For transcript analysis,total RNA was isolated using a total RNA minikit (IB47232; IBI) according to the manufacturer’s instruc-tions. Total RNA (400 ng) was converted to cDNA using the Maxima H Minus first-strand cDNA synthesiskit (K1681; Thermo Fisher) according to the manufacturer’s instructions. Quantitative reversetranscription-PCR (qRT-PCR) was performed as previously described (76) in technical triplicate. Meltingcurve analysis confirmed the presence of a single PCR product for each primer set. Data were analyzedwith SDS 2.4 software (Applied Biosystems), and viral gene expression was normalized to cellular GAPDHusing the ΔCT method, where CT is threshold cycle (77). Primer sequences are provided in Table 1.

RLM-RACE. G0- or S-phase synchronized cells were infected with WT Towne for 6 h at an MOI of 3.RNA was isolated from 1 � 106 cells using a Total RNA minikit (IB47232; IBI) according to the manufac-turer’s instructions. Isolated RNA was used directly for RLM-RACE using the GeneRacer kit with Super-Script III RT and TOPO TA cloning kit for sequencing (L150201; Invitrogen), following the manufacturer’sinstructions.

Western blotting. Cells were lysed in 1% SDS containing 2% �-mercaptoethanol and boiled for20 min prior to separation by SDS-PAGE and transfer to Optitran membranes (GE Healthcare). Mem-branes were blocked in 5% bovine serum albumin (BSA) in Tris-buffered saline with Tween 20 (TBST)

Lyon et al. ®

September/October 2020 Volume 11 Issue 5 e02410-20 mbio.asm.org 14

on February 28, 2021 by guest

http://mbio.asm

.org/D

ownloaded from

followed by incubations in primary and secondary antibody (10). Membranes were washed in TBST andimaged with the Odyssey Fc imager (Li-Cor). Images were generated with Image Studio version 5.2software (Li-Cor).

Data presentation and analysis. All graphs present means and standard deviations for threebiological replicates. An unpaired, two-tailed Student’s t test was used to determine the statisticaldifference between G0 and S samples where indicated.

ACKNOWLEDGMENTSThis work was supported by NIH grants AI139180 and AI130089 to R.F.K. S.M.L. was

supported by NRSA award T32 CA009135. The funders had no role in study design, datacollection and interpretation, or the decision to submit the work for publication.

We thank Jeff Meier and Mark Stinski for providing research materials, Kalejta labmembers for thoughtful discussions, and Emily Albright for comments on the manu-script.

S.M.L. performed all the experiments except for the coimmunoprecipitations, whichwere performed by K.D.Y, and the generation and characterization of the ΔCTD virus,which were performed by C.P. and M.N. S.M.L., K.D.Y., and R.F.K. designed experiments,analyzed the data, and wrote the manuscript.

REFERENCES1. Nascimento R, Costa H, Parkhouse RME. 2012. Virus manipulation of

cell cycle. Protoplasma 249:519 –528. https://doi.org/10.1007/s00709-011-0327-9.

2. Bresnahan WA, Boldogh I, Thompson EA, Albrecht T. 1996. Humancytomegalovirus inhibits cellular DNA synthesis and arrests productivelyinfected cells in late G1. Virology 224:150 –160. https://doi.org/10.1006/viro.1996.0516.

3. Dittmer D, Mocarski ES. 1997. Human cytomegalovirus infection inhibitsG1/S transition. J Virol 71:1629 –1634. https://doi.org/10.1128/JVI.71.2.1629-1634.1997.

4. Kalejta RF, Shenk T. 2003. The human cytomegalovirus UL82 gene product(pp71) accelerates progression through the G1 phase of the cell cycle. J Virol77:3451–3459. https://doi.org/10.1128/jvi.77.6.3451-3459.2003.

5. Kalejta RF, Bechtel JT, Shenk T. 2003. Human cytomegalovirus pp71stimulates cell cycle progression by inducing the proteasome-dependent degradation of the retinoblastoma family of tumor suppres-sors. Mol Cell Biol 23:1885–1895. https://doi.org/10.1128/mcb.23.6.1885-1895.2003.

6. Lu M, Shenk T. 1996. Human cytomegalovirus infection inhibits cellcycle progression at multiple points, including the transition from G1to S. J Virol 70:8850 – 8857. https://doi.org/10.1128/JVI.70.12.8850-8857.1996.

7. Jault FM, Jault JM, Ruchti F, Fortunato EA, Clark C, Corbeil J, Richman DD,Spector DH. 1995. Cytomegalovirus infection induces high levels ofcyclins, phosphorylated Rb, and p53, leading to cell cycle arrest. J Virol69:6697– 6704. https://doi.org/10.1128/JVI.69.11.6697-6704.1995.

8. Hume AJ, Finkel JS, Kamil JP, Coen DM, Culbertson MR, Kalejta RF.2008. Phosphorylation of retinoblastoma protein by viral protein withcyclin-dependent kinase function. Science 320:797–799. https://doi.org/10.1126/science.1152095.

9. Wiebusch L, Hagemeier C. 1999. Human cytomegalovirus 86-kilodaltonIE2 protein blocks cell cycle progression in G1. J Virol 73:9274 –9283.https://doi.org/10.1128/JVI.73.11.9274-9283.1999.

10. Saffert RT, Kalejta RF. 2006. Inactivating a cellular intrinsic immunedefense mediated by Daxx is the mechanism through which the humancytomegalovirus pp71 protein stimulates viral immediate-early gene

TABLE 1 Oligonucleotides used in this study

Primer Sequence Reference

qPCR primersIE19 forward 5=-GCAGAACTCGAGTCCCCT-3= This paperIE19 reverse 5=-TTACTGGTCAGCCTTGCT-3= This paperExon 4 forward 5=-TATACCCAGACGGAAGAGAAAT-3= 39Exon 4 reverse 5=-CCTTCAGTGCACCCCCTAACTT-3= 39Exon 3-4 (IE1) forward 5’-TTCCCAGAATTGGCCGAAGAA-3= This paperExon 3-4 (IE1) reverse 5’-CGCACCATGTCCACTCGAAC-3= This paper5’ RACE GSP1 CTD reverse 5=-CTGGTCAGCCTTGCTTCTAGTCA-3= This paper5’ RACE GSP2 CTD reverse 5=-CTCAGCACCATCCTCCTCTTCCT-3= This paperGAPDH forward 5=-GAGCCAAAAGGGTCATC-3’ 75GAPDH reverse 5=-GTGGTCATGAGTCCTTC-3= 75Exon 3 forward 5=-CGACGTTCCTGCAGACTATG-3= 73Exon 3 reverse 5=-TCCTCGGTCACTTGTTCAAA-3= 73UL54 promoter forward 5=-CACCAAAGACACGTCGTT-3= 74UL54 promoter reverse 5=-GTCCTTTGCGACCAGAAT-3= 74

Cloning primersFLAG-IE1/IE19 forward 5=-CTCACTATAGGGCGAATTCATGGACTACAAAGACCATGACGGTGATTATAAAGATCATG

ACATCGATTACAAGGATGACGATGACAAGGAGTCCTCTGCC-3’This paper

FLAG-IE1/IE1x4/IE19 reverse 5=-TTTAATAAGATCTGGATCCTTACTGGTCAGCCTTGCT-3= This paperFLAG-IE1x4 forward 5=-CTCACTATAGGGCGAATTCATGGACTACAAAGACCATGACGGTGATTATAAAGATCATG

ACATCGATTACAAGGATGACGATGACAAGGTGCGGCATAGAATCAA-3’This paper

FLAG-IE19ΔCTD reverse 5=-TTTAATAAGATCTGGATCCTTAAGAGGCGGTGGGTTC-3= This paper

IE19 Helps the HCMV Genome through Mitosis ®

September/October 2020 Volume 11 Issue 5 e02410-20 mbio.asm.org 15

on February 28, 2021 by guest

http://mbio.asm

.org/D

ownloaded from

expression. J Virol 80:3863–3871. https://doi.org/10.1128/JVI.80.8.3863-3871.2006.

11. Cantrell SR, Bresnahan WA. 2006. Human cytomegalovirus (HCMV) UL82gene product (pp71) relieves hDaxx-mediated repression of HCMV rep-lication. J Virol 80:6188 – 6191. https://doi.org/10.1128/JVI.02676-05.

12. Fortunato EA, Sanchez V, Yen JY, Spector DH. 2002. Infection of cells withhuman cytomegalovirus during S phase results in a blockade toimmediate-early gene expression that can be overcome by inhibition ofthe proteasome. J Virol 76:5369 –5379. https://doi.org/10.1128/jvi.76.11.5369-5379.2002.

13. Salvant BS, Fortunato EA, Spector DH. 1998. Cell cycle dysregulation byhuman cytomegalovirus: influence of the cell cycle phase at the time ofinfection and effects on cyclin transcription. J Virol 72:3729 –3741.https://doi.org/10.1128/JVI.72.5.3729-3741.1998.

14. Hertel L, Chou S, Mocarski ES. 2007. Viral and cell cycle-regulated kinasesin cytomegalovirus-induced pseudomitosis and replication. PLoS Pathog3:e6. https://doi.org/10.1371/journal.ppat.0030006.

15. Weisbach H, Schablowsky C, Vetter B, Gruska I, Hagemeier C, WiebuschL. 2017. Synthetic lethal mutations in the cyclin A interface of humancytomegalovirus. PLoS Pathog 13:e1006193. https://doi.org/10.1371/journal.ppat.1006193.

16. Bogdanow B, Weisbach H, Von Einem J, Straschewski S, Voigt S, WinklerM, Hagemeier C, Wiebusch L. 2013. Human cytomegalovirus tegumentprotein pp150 acts as a cyclin A2-CDK-dependent sensor of the host cellcycle and differentiation state. Proc Natl Acad Sci U S A 110:17510 –17515. https://doi.org/10.1073/pnas.1312235110.

17. Vogt VM. 1997. Retroviral virions and genomes. In Coffin J, Hughes S,Varmus H (ed), Retroviruses. Cold Spring Harbor Laboratory Press, ColdSpring Harbor, NY.

18. Aydin I, Schelhaas M. 2016. Viral genome tethering to host cellchromatin: cause and consequences. Traffic 17:327–340. https://doi.org/10.1111/tra.12378.

19. Chiu Y-F, Sugden B. 2018. Plasmid partitioning by human tumor viruses.J Virol 92:e02170-17. https://doi.org/10.1128/JVI.02170-17.

20. Coursey TL, McBride AA. 2019. Hitchhiking of viral genomes on cellularchromosomes. Annu Rev Virol 6:275–296. https://doi.org/10.1146/annurev-virology-092818-015716.

21. Piirsoo M, Ustav E, Mandel T, Stenlund A, Ustav M. 1996. Cis and transrequirements for stable episomal maintenance of the BPV-1 replicator.EMBO J 15:1–11. https://doi.org/10.1002/j.1460-2075.1996.tb00328.x.

22. Kanda T, Otter M, Wahl GM. 2001. Coupling of mitotic chromosometethering and replication competence in Epstein-Barr virus-based plas-mids. Mol Cell Biol 21:3576 –3588. https://doi.org/10.1128/MCB.21.10.3576-3588.2001.

23. Ballestas ME, Kaye KM. 2001. Kaposi’s sarcoma-associated herpesviruslatency-associated nuclear antigen 1 mediates episome persistencethrough cis-acting terminal repeat (TR) sequence and specificallybinds TR DNA. J Virol 75:3250 –3258. https://doi.org/10.1128/JVI.75.7.3250-3258.2001.

24. Cotter MA, Subramanian C, Robertson ES. 2001. The Kaposi’s sarcoma-associated herpesvirus latency-associated nuclear antigen binds to spe-cific sequences at the left end of the viral genome through its carboxy-terminus. Virology 291:241–259. https://doi.org/10.1006/viro.2001.1202.

25. Fejér G, Medveczky MM, Horvath E, Lane B, Chang Y, Medveczky PG.2003. The latency-associated nuclear antigen of Kaposi’s sarcoma-associated herpesvirus interacts preferentially with the terminal re-peats of the genome in vivo and this complex is sufficient forepisomal DNA replication. J Gen Virol 84:1451–1462. https://doi.org/10.1099/vir.0.18940-0.

26. Skiadopoulos MH, McBride AA. 1998. Bovine papillomavirus type 1genomes and the E2 transactivator protein are closely associated withmitotic chromatin. J Virol 72:2079 –2088. https://doi.org/10.1128/JVI.72.3.2079-2088.1998.

27. You J, Croyle JL, Nishimura A, Ozato K, Howley PM. 2004. Interaction ofthe bovine papillomavirus E2 protein with Brd4 tethers the viral DNA tohost mitotic chromosomes. Cell 117:349 –360. https://doi.org/10.1016/s0092-8674(04)00402-7.

28. Parish JL, Bean AM, Park RB, Androphy EJ. 2006. ChlR1 is required forloading papillomavirus E2 onto mitotic chromosomes and viral ge-nome maintenance. Mol Cell 24:867– 876. https://doi.org/10.1016/j.molcel.2006.11.005.

29. Sears J, Ujihara M, Wong S, Ott C, Middeldorp J, Aiyar A. 2004. The aminoterminus of Epstein-Barr virus (EBV) nuclear antigen 1 contains AT hooksthat facilitate the replication and partitioning of latent EBV genomes by

tethering them to cellular chromosomes. J Virol 78:11487–11505.https://doi.org/10.1128/JVI.78.21.11487-11505.2004.

30. Chakravorty A, Sugden B. 2015. The AT-hook DNA binding ability of theEpstein Barr virus EBNA1 protein is necessary for the maintenance ofviral genomes in latently infected cells. Virology 484:251–258. https://doi.org/10.1016/j.virol.2015.05.018.

31. Barbera AJ, Chodaparambil JV, Kelley-Clarke B, Joukov V, Walter JC,Luger K, Kaye KM. 2006. The nucleosomal surface as a docking station forKaposi’s sarcoma herpesvirus LANA. Science 311:856 – 861. https://doi.org/10.1126/science.1120541.

32. Barbera AJ, Chodaparambil JV, Kelley-Clarke B, Luger K, Kaye KM.2006. Kaposi’s sarcoma-associated herpesvirus LANA hitches a rideon the chromosome. Cell Cycle 5:1048 –1052. https://doi.org/10.4161/cc.5.10.2768.

33. Lafemina RL, Pizzorno MC, Mosca JD, Hayward GS. 1989. Expressionof the acidic nuclear immediate-early protein (IE1) of human cyto-megalovirus in stable cell lines and its preferential association withmetaphase chromosomes. Virology 172:584 – 600. https://doi.org/10.1016/0042-6822(89)90201-8.

34. Wilkinson GWG, Kelly C, Sinclair JH, Rickards C. 1998. Disruption ofPML-associated nuclear bodies mediated by the human cytomegalovi-rus major immediate early gene product. J Gen Virol 79:1233–1245.https://doi.org/10.1099/0022-1317-79-5-1233.

35. Shin HJ, Kim YE, Kim ET, Ahn JH. 2012. The chromatin-tethering domainof human cytomegalovirus immediate-early (IE) 1 mediates associationsof IE1, PML and STAT2 with mitotic chromosomes, but is not essential forviral replication. J Gen Virol 93:716 –721. https://doi.org/10.1099/vir.0.037986-0.

36. Reinhardt J, Smith GB, Himmelheber CT, Azizkhan-Clifford J, Mocarski ES.2005. The carboxyl-terminal region of human cytomegalovirus IE1491aacontains an acidic domain that plays a regulatory role and a chromatin-tethering domain that is dispensable during viral replication. J Virol79:225–233. https://doi.org/10.1128/JVI.79.1.225-233.2005.

37. Mücke K, Paulus C, Bernhardt K, Gerrer K, Schön K, Fink A, Sauer E-M,Asbach-Nitzsche A, Harwardt T, Kieninger B, Kremer W, Kalbitzer HR,Nevels M. 2014. Human cytomegalovirus major immediate early 1 pro-tein targets host chromosomes by docking to the acidic pocket on thenucleosome surface. J Virol 88:1228 –1248. https://doi.org/10.1128/JVI.02606-13.

38. Fang Q, Chen P, Wang M, Fang J, Yang N, Li G, Xu R-M. 2016. Humancytomegalovirus IE1 protein alters the higher-order chromatin structureby targeting the acidic patch of the nucleosome. Elife 5:1–11. https://doi.org/10.7554/eLife.11911.

39. Tarrant-Elorza M, Rossetto CC, Pari GS. 2014. Maintenance and replica-tion of the human cytomegalovirus genome during latency. Cell HostMicrobe 16:43–54. https://doi.org/10.1016/j.chom.2014.06.006.

40. Mauch-Mücke K, Schön K, Paulus C, Nevels MM. 2020. Evidence fortethering of human cytomegalovirus genomes to host chromosomes.Front Cell Infect Microbiol. https://doi.org/10.3389/fcimb.2020.577428.

41. Scherer M, Klingl S, Sevvana M, Otto V, Schilling E, Reuter N, Sticht H,Muller YA, Stump JD, Mu R, Stamminger T. 2014. Crystal structure ofcytomegalovirus IE1 protein reveals targeting of TRIM family memberPML via coiled-coil interactions. PLoS Pathog 10:e1004512. https://doi.org/10.1371/journal.ppat.1004512.

42. Stump JD, Sticht H. 2016. Investigation of the dynamics of the viralimmediate- early protein 1 in different conformations and oligomeriza-tion states. J Biomol Struct Dyn 34:1029 –1041. https://doi.org/10.1080/07391102.2015.1065204.

43. Barry PA, Alcendor DJ, Power MD, Kerr H, Luciw PA. 1996. Nucleotidesequence and molecular analysis of the rhesus cytomegalovirusimmediate-early gene and the UL121-117 open reading frames. Virology215:61–72. https://doi.org/10.1006/viro.1996.0007.

44. Qin Q, Penkert RR, Kalejta RF. 2013. Heterologous viral promoters incor-porated into the human cytomegalovirus genome are silenced duringexperimental latency. J Virol 87:9886 –9894. https://doi.org/10.1128/JVI.01726-13.

45. Chiu Y-F, Sugden AU, Fox K, Hayes M, Sugden B. 2017. Kaposi’s sarcoma-associated herpesvirus stably clusters its genomes across generations tomaintain itself extrachromosomally. J Cell Biol 216:2745–2758. https://doi.org/10.1083/jcb.201702013.

46. Nanbo A, Sugden A, Sugden B. 2007. The coupling of synthesis andpartitioning of EBV’s plasmid replicon is revealed in live cells. EMBO J26:4252– 4262. https://doi.org/10.1038/sj.emboj.7601853.

47. You J, Schweiger M-R, Howley PM. 2005. Inhibition of E2 binding to Brd4

Lyon et al. ®

September/October 2020 Volume 11 Issue 5 e02410-20 mbio.asm.org 16

on February 28, 2021 by guest

http://mbio.asm

.org/D

ownloaded from

enhances viral genome loss and phenotypic reversion of bovinepapillomavirus-transformed cells. J Virol 79:14956 –14961. https://doi.org/10.1128/JVI.79.23.14956-14961.2005.

48. Gurczynski SJ, Das S, Pellett PE. 2014. Deletion of the human cytomeg-alovirus US17 gene increases the ratio of genomes per infectious unitand alters regulation of immune and endoplasmic reticulum stressresponse genes at early and late times after infection. J Virol 88:2168 –2182. https://doi.org/10.1128/JVI.02704-13.

49. Awasthi S, Isler JA, Alwine JC. 2004. Analysis of splice variants of theimmediate-early 1 region of human cytomegalovirus. J Virol 78:8191– 8200. https://doi.org/10.1128/JVI.78.15.8191-8200.2004.

50. Du G, Dutta N, Lashmit P, Stinski MF. 2011. Alternative splicing of thehuman cytomegalovirus major immediate-early genes affects infectious-virus replication and control of cellular cyclin-dependent kinase. J Virol85:804 – 817. https://doi.org/10.1128/JVI.01173-10.

51. Shirakata M, Terauchi M, Ablikim M, Imadome K-I, Hirai K, Aso T, Yama-nashi Y. 2002. Novel immediate-early protein IE19 of human cytomeg-alovirus activates the origin recognition complex I promoter in a coop-erative manner with IE72. J Virol 76:3158 –3167. https://doi.org/10.1128/jvi.76.7.3158-3167.2002.

52. Bochkarev A, Barwell JA, Pfuetzner RA, Furey W, Edwards AM, Frappier L,Ln C. 1995. Crystal structure of the DNA-binding of the Epstein-Barr virusorigin-binding protein EBNA1. Cell 83:39 – 46. https://doi.org/10.1016/0092-8674(95)90232-5.

53. Cardenas-Mora J, Spindler JE, Jang MK, McBride AA. 2008. Dimerizationof the papillomavirus E2 protein is required for efficient mitotic chro-mosome association and Brd4 binding. J Virol 82:7298 –7305. https://doi.org/10.1128/JVI.00772-08.

54. Chen MR, Middeldorp JM, Hayward SD. 1993. Separation of the complexDNA binding domain of EBNA-1 into DNA recognition and dimerizationsubdomains of novel structure. J Virol 67:4875– 4885. https://doi.org/10.1128/JVI.67.8.4875-4885.1993.

55. Hegde RS, Grossman SR, Laimins LA, Sigler PB. 1992. Crystal structure at1.7 Å of the bovine papillomavirus-1 E2 DMA-binding domain bound toits DNA target. Nature 359:505–512. https://doi.org/10.1038/359505a0.

56. Komatsu T, Ballestas ME, Barbera AJ, Kelley-Clarke B, Kaye KM. 2004.KSHV LANA1 binds DNA as an oligomer and residues N-terminal to theoligomerization domain are essential for DNA binding, replication, andepisome persistence. Virology 319:225–236. https://doi.org/10.1016/j.virol.2003.11.002.

57. Matagne A, Joris B, Frere JM. 1991. Anomalous behaviour of a proteinduring SDS/PAGE corrected by chemical modification of carboxylicgroups. Biochem J 280:553–556. https://doi.org/10.1042/bj2800553.

58. Armstrong DJ, Roman A. 1993. The anomalous electrophoretic behaviorof the human papillomavirus type 16 E7 protein is due to the highcontent of acidic amino acid residues. Biochem Biophys Res Commun192:1380 –1387. https://doi.org/10.1006/bbrc.1993.1569.

59. Shi W, Huang Y, Sutton-Smith M, Tissot B, Panico M, Morris HR, Dell A,Haslam SM, Boyington J, Graham BS, Yang Z-Y, Nabel GJ. 2008. Afilovirus-unique region of Ebola virus nucleoprotein confers aberrantmigration and mediates its incorporation into virions. J Virol 82:6190 – 6199. https://doi.org/10.1128/JVI.02731-07.

60. Kern F, Surel IP, Faulhaber N, Frömmel C, Schneider-Mergener J, Schöne-mann C, Reinke P, Volk H-D. 1999. Target structures of the CD8�-T-cellresponse to human cytomegalovirus: the 72-kilodalton majorimmediate-early protein revisited. J Virol 73:8179 – 8184. https://doi.org/10.1128/JVI.73.10.8179-8184.1999.

61. Cobbs CS, Harkins L, Samanta M, Gillespie GY, Bharara S, King PH, Nabors

LB, Cobbs CG, Britt WJ. 2002. Human cytomegalovirus infection andexpression in human malignant glioma. Cancer Res 62:3347–3350.

62. Dziurzynski K, Chang SM, Heimberger AB, Kalejta RF, Dallas SRMG, SmitM, Soroceanu L, Cobbs CS, HCMV and Gliomas Symposium. 2012. Con-sensus on the role of human cytomegalovirus in glioblastoma. NeuroOncol 14:246 –255. https://doi.org/10.1093/neuonc/nor227.

63. Ranganathan P, Clark PA, Kuo JS, Salamat MS, Kalejta RF. 2012. Signifi-cant association of multiple human cytomegalovirus genomic loci withglioblastoma multiforme samples. J Virol 86:854 – 864. https://doi.org/10.1128/JVI.06097-11.

64. Liu C, Clark PA, Kuo JS, Kalejta RF. 2017. Human cytomegalovirus-infected glioblastoma cells display stem cell-like phenotypes. mSphere2:e00137-17. https://doi.org/10.1128/mSphere.00137-17.

65. Chen R, Xiong S, Yang Y, Fu W, Wang Y, Ge J. 2003. The relationshipbetween human cytomegalovirus infection and atherosclerosis de-velopment. Mol Cell Biochem 249:91–96. https://doi.org/10.1023/A:1024759311013.

66. Cheng J, Ke Q, Jin Z, Wang H, Kocher O, Morgan JP, Zhang J, CrumpackerCS. 2009. Cytomegalovirus infection causes an increase of arterial bloodpressure. PLoS Pathog 5:e1000427. https://doi.org/10.1371/journal.ppat.1000427.

67. Griffiths P, Baraniak I, Reeves M. 2015. The pathogenesis of humancytomegalovirus. J Pathol 235:288 –297. https://doi.org/10.1002/path.4437.

68. Sinclair J. 2010. Chromatin structure regulates human cytomegalovirusgene expression during latency, reactivation and lytic infection. BiochimBiophys Acta 1799:286 –295. https://doi.org/10.1016/j.bbagrm.2009.08.001.

69. Penkert RR, Kalejta RF. 2011. Tegument protein control of latent herpes-virus establishment and animation. Herpesviridae 2:3. https://doi.org/10.1186/2042-4280-2-3.

70. Zhu H, Shen Y, Shenk T. 1995. Human cytomegalovirus IE1 and IE2proteins block apoptosis. J Virol 69:7960 –7970. https://doi.org/10.1128/JVI.69.12.7960-7970.1995.

71. Albright ER, Kalejta RF. 2016. Canonical and variant forms of histone H3are deposited onto the human cytomegalovirus genome during lyticand latent infections. J Virol 90:10309 –10320. https://doi.org/10.1128/JVI.01220-16.

72. Paulus C, Harwardt T, Walter B, Marxreiter A, Zenger M, Reuschel E,Nevels MM. 2020. Revisiting promyelocytic leukemia protein targetingby human cytomegalovirus immediate-early protein 1. PLoS Pathog16:e1008537. https://doi.org/10.1371/journal.ppat.1008537.

73. Hwang J, Saffert RT, Kalejta RF. 2011. Elongin B-mediated epigeneticalteration of viral chromatin correlates with efficient human cytomega-lovirus gene expression and replication. mBio 2:e00023-11. https://doi.org/10.1128/mBio.00023-11.

74. Nitzsche A, Paulus C, Nevels M. 2008. Temporal dynamics of cytomeg-alovirus chromatin assembly in productively infected human cells. J Virol82:11167–11180. https://doi.org/10.1128/JVI.01218-08.

75. Juckem LK, Boehme KW, Feire AL, Compton T. 2008. Differential initia-tion of innate immune responses induced by human cytomegalovirusentry into fibroblast cells. J Immunol 180:4965– 4977. https://doi.org/10.4049/jimmunol.180.7.4965.

76. Lee SH, Albright ER, Lee J-H, Jacobs D, Kalejta RF. 2015. Cellular defenseagainst latent colonization foiled by human cytomegalovirus UL138protein. Sci Adv 1:e1501164. https://doi.org/10.1126/sciadv.1501164.

77. Schmittgen TD, Livak KJ. 2008. Analyzing real-time PCR data by thecomparative C(T) method. Nat Protoc 3:1101–1108. https://doi.org/10.1038/nprot.2008.73.

IE19 Helps the HCMV Genome through Mitosis ®

September/October 2020 Volume 11 Issue 5 e02410-20 mbio.asm.org 17

on February 28, 2021 by guest

http://mbio.asm

.org/D

ownloaded from